Coordinated Lane-Level Variable Speed Limits and Ramp Metering for Successive Weaving Segments Considering Merging/Diverging Risks: A Hybrid Model Predictive Control and Multi-Agent Reinforcement Learning Approach
A hybrid MPC and multi-agent RL controller coordinates lane-level speed limits and ramp metering in Changchun.
The paper proposes a hybrid model-predictive-control and multi-agent reinforcement-learning framework for coordinating lane-level variable speed limits and ramp metering on successive weaving segments. L-METANET models lane-level traffic, while XGBoost-SHAP and a random-parameters logit model estimate merging and diverging collision risk. Experiments on the 18-km Eastern Expressway in Changchun, China, show faster convergence than MPC and MARL baselines and better generalization than pure MARL under fluctuating demand.
- Hybrid MPC and multi-agent RL coordinates speed limits and ramp meters
- Lane-level L-METANET captures free and forced lane changes
- Risk models exceed 0.80 AUC on most tasks
- Tested on Changchun's 18-km Eastern Expressway
- Tags
- reinforcement-learningmulti-agentmodel-predictive-controltraffic-managementvariable-speed-limitsramp-metering
- Products
- L-METANETMPC-STMAPPO
- Countries
- China
From the original article
1 min read at arxiv.orgSuccessive weaving segments (SWSs) on urban expressways are bottlenecks prone to recurrent congestion and collisions, requiring fine-grained active traffic management (ATM). Existing approaches struggle to balance the adaptive performance of data-driven optimization with the resilience and transferability of model-based control.
By Guodong Ma · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.